A Contrastive Distillation Approach for Incremental Semantic Segmentation in Aerial Images

TL;DR

This paper introduces a contrastive distillation method to improve incremental semantic segmentation in aerial images, addressing catastrophic forgetting by leveraging orientation-based contrastive regularization, and demonstrates superior performance on the Potsdam dataset.

Contribution

It proposes a novel contrastive regularization technique combined with knowledge distillation for incremental semantic segmentation in aerial imagery, considering orientation features unique to aerial data.

Findings

Outperforms baseline methods on the Potsdam dataset

Effectively mitigates catastrophic forgetting in incremental segmentation

Utilizes orientation-based contrastive regularization for improved feature retention

Abstract

Incremental learning represents a crucial task in aerial image processing, especially given the limited availability of large-scale annotated datasets. A major issue concerning current deep neural architectures is known as catastrophic forgetting, namely the inability to faithfully maintain past knowledge once a new set of data is provided for retraining. Over the years, several techniques have been proposed to mitigate this problem for image classification and object detection. However, only recently the focus has shifted towards more complex downstream tasks such as instance or semantic segmentation. Starting from incremental-class learning for semantic segmentation tasks, our goal is to adapt this strategy to the aerial domain, exploiting a peculiar feature that differentiates it from natural images, namely the orientation. In addition to the standard knowledge distillation approach, we propose a contrastive regularization, where any given input is compared with its augmented version (i.e. flipping and rotations) in order to minimize the difference between the segmentation features produced by both inputs. We show the effectiveness of our solution on the Potsdam dataset, outperforming the incremental baseline in every test. Code available at: https://github.com/edornd/contrastive-distillation.